# Forecasting more than one future.

GenCast used machine learning to generate ensembles of possible weather outcomes up to 15 days ahead.

Canonical: https://brightaifuture.com/discoveries/gencast
Format: discovery
Source publication: 2024-12-04
Bright publication: 2026-09-05
Substantive update: None recorded
Evidence and review: Demonstrated; confidence: unassessed; source-checked; legacy source check. Legacy source check; no named human reviewer is recorded in this projection.

## The human problem

A single forecast does not communicate the range of weather that could occur.

## The prior constraint

Ensemble forecasting uses multiple simulations to represent uncertainty.

## AI’s actual role

A diffusion model generated probabilistic global forecasts.

## The documented result

The research reported stronger skill than the comparison operational ensemble on the evaluated targets.

## Why it may matter

Better uncertainty estimates could support weather-dependent decisions.

## Limitations

Retrospective evaluation is not proof of operational benefit in every region or extreme event.

## Unresolved questions

Test operational reliability and local decision usefulness.

## Provenance and history

{
  "dates": {
    "eventDate": null,
    "publicationDate": "2024-12-04",
    "captureDate": null,
    "lastReviewedDate": "2026-09-05"
  },
  "provenance": {
    "origin": "legacy-projection",
    "externalId": "gencast"
  },
  "revisions": [],
  "corrections": []
}

## Original sources

- [Probabilistic weather forecasting with machine learning · Nature](https://www.nature.com/articles/s41586-024-08252-9)

## Continue exploring

- [Planet](https://brightaifuture.com/worlds/planet)
- [How much earlier could we understand what is coming?](https://brightaifuture.com/threads/weather)
- [Extended story and historical record](https://brightaifuture.com/stories/forecasting-more-than-one-future)
- [Open this record in the interactive world](https://brightaifuture.com/?event=gencast)